International Journal of Engineering Inventions

e-ISSN: 2278-7461, p-ISSN: 2319-6491

Volume 2, Issue 8 (May 2013) PP: 16-27

Evaluation of Engineering Stresses as the “Correct” Measure of

“Physical” Stresses in Large Strain Geometrically Nonlinear
Problems
A. E. Mohmed1, N. M. Akasha2, F. M. Adam3
1
Dept. of Civil Engineering, Sudan University of Science and Technology, Sudan
2
Dept. of Civil Engineering, Sudan University of Science and Technology, Sudan
3
Dept. of Civil Engineering, Jazan University, Kingdom of Saudi Arabia

Abstract: In this paper, Lagrangian formulations for geometric nonlinear plane stress/strain problems based on
different stress measures are evaluated. A Total Lagrangian formulation based on the exact Engineering strains
is developed. The 2ndPiola-Kirchhoff stresses based on the well known Green strains and the Engineering
stresses based on the exact Engineering (geometric or conventional) strains obtained from Total Lagrangian
formulations are compared with the true Cauchy stresses. The Engineering stresses based on the assumption of
small shear strains are also compared with the above mentioned stresses. Geometric nonlinear Total
Lagrangian formulations applied on two-dimensional elasticity using 4-node plane finite elements were used.
The formulations were implemented into the finite element program (NUSAP). The solution of nonlinear
equations was obtained by the Newton-Raphson method. The program was applied to obtain stresses for the
different strain measures. The true Cauchy stresses were obtained by using the Logarithmic strains. The
evaluation was based on comparing the results of three numerical examples. For moderate and large strains,
the exact Engineering stresses are good measures of the correct physical stresses. Thus, these must be used
when the stresses are required from a Total Lagrangian solution.
Keywords: Cauchy stresses, Engineering strain, Geometric Nonlinear, Plain stress/strain, Total Lagrangian

I. INTRODUCTION
Almost all structures behave in some nonlinear manner prior to reaching their limit of resistance. For
this reason, most modern codes have incorporated certain provisions to consider nonlinear effects e.g. limit state
design methods. Also, the use of light, high strength materials, resulting in light “tall” structures, introduces
certain degrees of nonlinearity. This coupled with advance in solution methods and computing facilities make
room for geometric nonlinear analysis. The major problem in geometrically nonlinear (GNL) finite element
analysis is the need to define reference coordinates and to specify the relevant stress and strain measures.
The two main finite element formulations for GNL problems are the Eulerian formulations (EFMS)
and the Lagrangian formulations (LFMS). As stated, among others, by Yang and Kuo [1], Crisfield [2],
Zeinkiewicz and Taylor [3] and Mohamed [4] LFMS, in contrast to EFMS, are suitable in solid mechanics
applications. This is mainly due to the ease with which they handle complicated boundaries and their ability to
follow material points enabling the accurate treatment of history dependent materials. There are two main
approaches to LFMS, namely the Total Lagrangian (TL) and the Updated Lagrangian (UL). Yang and Kuo [1],
Zeinkiewicz and Taylor [3], Belytschko [5] and Marinkowic et al [6], stated that the UL approach provides the
most efficient formulation and can be considered equivalent to the EFM. Wood and Zeinkiewicz [7] stated that
the TL approach offers advantages since the initial configuration remains constant which simplifies formulation
and computation. Surana and Sorem [8] presented a TL approach for framed structures with no restrictions on
rotation increments. Djermane et al [9] pointed out that the TL formulation is now recognized as the most
realistic civil engineering approach. But, the main serious drawback of the TL approach is that it is based on the
Green strains, which are unsuitable for work with large strains. Also, the 2ndPiola-Kirchhoff stresses, which are
work conjugate to the Green strains, are defined in the deformed configuration and should be transformed to the
undeformed configuration [1], [2], [3], [4] and [8]. Thus, the TL approach while being very well established,
having the above mentioned advantages and giving accurate displacement values, will result in stresses with no
physical significance [2].
As for the definition of stress and strain measures, Crisfield [2] and Bonet and Wood [10] proposed the
use, as work conjugate in the virtual work expression, the Green strains with the 2ndPiola-Kirchhoff stresses, the
Engineering (conventional or geometric) strains with the Engineering stresses and the Logarithmic strains with
the Cauchy "true" stresses. Yang and Kuo [1] used, as work conjugate, the Green-Lagrange strains with the
2ndPiola-Kirchhoff stresses, the Infinitesimal strains with the Cauchy stresses and the Updated Green strains
with the Updated Kirchhoff stresses. They adopted the Updated Kirchhoff stresses as the measure of the "true"
www.ijeijournal.com Page | 16
Evaluation of Engineering Stresses as the “Correct” Measure of “Physical” Stresses in Large Strain

physical stresses. In fact, the proposition that: using the UL approach converts the 2ndPiola-Kirchhoff stresses to
the "true" stresses is also adopted by many researchers [2], [3], [5], [6] and [10].
An alternative TL approach that results in evaluating "correct" stresses was developed by Surana and
Sorem [8]. Their formulation for framed structures removed the restrictions of small rotation increments by
retaining nonlinear terms in the definition of the element displacement field. Mohamed [4] presented a TL
formulation based on the Engineering strains which resulted in the correct stresses for small strain large rotation
deformation of beams. Mohamed and Adam [11], [12] extended the TL formulation based on Engineering
strains to the analysis of shell structures. Akasha [13] and Akasha and Mohamed [14], [15] developed a similar
TL formulation for plane stress/strain problems. These formulations were based on using the Engineering strains
with the Engineering stresses in the virtual work expression. The formulations were developed basing the
variation of the Engineering strains on the variation of Green strains. These formulations are similar to the
positional formulation based on Engineering strains used by Greco and Ferreira [16]. The formulations were,
also, extended to include the evaluation of the true Cauchy stresses based on Logarithmic strains [13], [14] and
[17]. The only limitation of these formulations is the assumption that the shear strains are small. However, it is
possible to avoid this limitation by considering the exact variation of the Engineering strains.
This paper presents a TL formulation based the exact variation of Engineering strains. The paper is,
also take an evaluation of the TL nonlinear stresses which are compared with the true Cauchy stresses presented
by Akasha and Mohamed [17] using the Logarithmic strains. The comparison is carried out for plane
stress/strain problems with the intention of investigating the effect of large strains on the different stress
measures. The effect of avoiding the limitation on shear strains is, also, looked into. The comparison is
considered as a criterion for ensuring the accuracy, consistency and convergence to the correct results of the
nonlinear analysis. Thus, the aim of the evaluation is pointing out when the TL formulation based on
Engineering strains is to be used, with confidence, in the analysis of large strain problems.

II. NONLINEAR FINITE ELEMENT FORMULATION

In the following sections there are two alternative Lagrangian formulations of the incremental
equilibrium equations for large strain two dimensional problems which are outlined below.

2.1 Geometrically Non-linear Finite Element TL Formulation based on Green strains (TLG):
The 2ndPiola-Kirchhoff stresses, work conjugate to the Green-Lagrange strains, are the internal forces
per unit initial area acting along the normal and tangential directions at the deformed configuration [1], [2] and
[4]. Thus, these stresses are referred to the convected coordinates in the deformed configuration but measured
per unit area of the undeformed body with attempts to take this into consideration in the virtual work expression
result in an unsymmetrical stiffness matrix [4]. Hence, direct proportionality between the 2ndPiola-Kirchhoff
stresses, s0 , and the Green-Lagrange strains, e0 , is assumed when writing the virtual work expression. In two
dimensions, with reference to the initial configuration (t=0), the Green strains are given by:
𝑇 1 T
𝒆𝟎 = 𝑒𝑥, 𝑒𝒚, 𝑒𝑥𝑦 = 𝐅 𝐅 − 𝐈 (1)
2
where F is the displacement gradient matrix.
In a finite element formulation equation (1) is written as:
1
𝒆𝟎 = 𝒆𝟎0 + 𝒆𝑳0 = 𝑩𝟎 𝒂𝟎 + 𝑩𝑳 𝒂𝟎 𝒂0 (2)
2
where 𝒂𝟎 is the vector of nodal variables.
The nonlinear strain eL0 can be written as:
1 1
𝒆𝑳𝟎 = 𝑩𝑳 𝒂𝟎 𝒂0 = 𝑨𝜃 𝑮0 𝒂0 (3)
2 2
The strain displacement matrix B is given by:
𝑩 = 𝑩𝟎 + 𝑩𝑳 𝒂𝟎 (4)
The tangent stiffness matrix now takes the form:
𝑲 𝑇 = 𝑲0 + 𝑲𝐿 𝒂 𝟎 + 𝑲𝜎 (5)
where:

𝑲0 + 𝑲𝐿 𝒂 𝟎 = 𝑩𝑇 𝑫𝑩 𝑑𝑉0 (6)𝑎
𝑉0
in which D is the modulus matrix, and the initial stress stiffness matrix is given by:

𝑲𝜎 = 𝑮𝑇0 𝑷0𝑖 𝑮0 𝑑𝑉0 (6)𝑏

𝑉0
in which 𝑷0𝑖 is the initial stress matrix.
The displacement increments ∆ai0 are evaluated by using K T and the residuals as:
www.ijeijournal.com Page | 17
Evaluation of Engineering Stresses as the “Correct” Measure of “Physical” Stresses in Large Strain

∆𝒂𝑖0 = − 𝑲−1 𝑇 𝝋
𝑖
(7)
The total displacements are, then, obtained as:
ai+1 i
0 = a 0 + ∆a 0
i
(8)
The strain increments are defined by:
1
∆𝒆𝑖𝑜 = 𝑩𝟎 + 𝑩𝑳 𝒂𝑖0 + 𝑩𝑳 ∆𝒂𝑖0 ∆𝒂𝑖0 (9)
2
And the stress increments are given by:
∆𝒔𝒊𝟎 = 𝑫∆𝒆𝑖𝑜 (10)
And the total stresses are:
𝒔𝒊+𝟏
𝟎 = 𝒔𝒊0 + ∆𝒔𝒊𝟎 (11)
From which the nodal residual forces are evaluated as follows:

−𝝋𝒊+𝟏 = 𝑹 − 𝑩𝑇 𝒔𝒊+𝟏
𝟎 𝑑𝑉0 (12)
𝑉0
where 𝑹 is the vector of applied equivalent nodal forces and:
𝑩 = 𝑩𝟎 + 𝑩𝑳 𝒂𝒊+𝟏 0

2.2 Geometrically Non-linear Finite Element TL Formulation based on Engineering strains (TLE):
In two dimensions the geometric strains, unit stretches, εx and εy are defined by the change in length
per unit initial length of line elements originally oriented parallel to the x and y axes respectively. The shear
strain γxy is the actual angle change.
The geometric strains, as defined above, are given in terms of Green strains by:
1 1
𝜀𝑥 = 𝑔𝑥 . 𝑔𝑥 2 − 1 = 1 + 2𝑒𝑥 2 − 1
1 1
𝜀𝑦 = 𝑔𝑦 . 𝑔𝑦 − 1 = 1 + 2𝑒𝑦 − 1
2 2 (13)
And the shear strain is defined as:
𝑒𝑥𝑦
𝛾𝑥𝑦 = sin−1 1 1 (14)
1 + 2𝑒𝑥 2 1 + 2𝑒𝑦 2
where:
1 1
𝑒𝑥𝑦 = 𝑔𝑥 . 𝑔𝑦 = 1 + 2𝑒𝑥 2 1 + 2𝑒𝑦 2 sin 𝛾𝑥𝑦
𝜕𝑅 𝜕𝑅
𝑔𝑥 = , 𝑔𝑦 = 𝜕𝑦 are the displacement gradient vectors, and R is the position vector after
𝜕𝑥
deformation.
The variation in the Engineering strains is given by:
𝛿𝜺0 = 𝑯 𝛿𝒆0 (15)
where:
𝛿𝒆0 = 𝑩0 + 𝑩𝐿 𝒂0 𝛿𝒂0
= 𝑩0 + 𝑨𝜃 𝑮0 𝛿𝒂0 = 𝑩𝛿𝒂0 (16)
From which, the variations in the Engineering strains are given by:
𝛿𝜺0 = 𝑯𝑩𝛿𝒂0 = 𝑩∗ 𝛿𝒂0 (17)
In which B is the strain matrix, and H relates variation in Engineering strains to variation in Green
strains.
The incremental equilibrium equations in terms of Engineering stresses are:

𝑲∗𝑇 ∆𝒂0 = 𝑹 − 𝑩𝑇 𝑯𝑇 𝝈 𝑑𝑉0 = 𝑹 − 𝑩∗𝑇 𝝈 𝑑𝑉0 (18)

𝑉0 𝑉0
where 𝝈 are the Engineering stresses.
The tangent stiffness matrix 𝑲∗𝑇 is now given by:
𝑲∗𝑇 = 𝑲∗0 + 𝑲∗𝐿 + 𝑲∗𝜎 + 𝑲∗∗ 𝜎 (19)
where:

𝑲∗0 = 𝑩𝑇0 𝑯𝑇 𝑫𝑯𝑩0 𝑑𝑉0 20 𝑎

𝑉0
and

𝑲∗𝐿 = 𝑩𝑇0 𝑯𝑇 𝑫𝑯𝑩𝐿 𝑑𝑉0 + 𝑩𝑇𝐿 𝑯𝑇 𝑫𝑯𝑩0 𝑑𝑉0 + 𝑩𝑇𝐿 𝑯𝑇 𝑫𝑯𝑩𝐿 𝑑𝑉0 20 𝑏
𝑉0 𝑉0 𝑉0
and 𝑲∗𝜎 is the symmetric matrix dependent on the Engineering stress, and can be written as:

www.ijeijournal.com Page | 18
Evaluation of Engineering Stresses as the “Correct” Measure of “Physical” Stresses in Large Strain

∗
𝑲∗𝜎 = 𝑮𝑇0 𝑷𝟎𝒊 𝑮0 𝑑𝑉0 (21)
𝑉0
where 𝑮0 is a matrix containing shape function derivatives.
∗
and the initial stress matrix 𝑷𝟎𝒊 is defined as:
∗ ∗
𝜎𝒙 𝐼 𝜎𝒙𝒚 𝐼
𝑷∗𝟎𝒊 = ∗ (22)
𝜎𝒙𝒚 𝐼 𝜎𝒚∗ 𝐼
where 𝐼 is 2  2 unit matrix.
and 𝝈∗ is the stress vector given by:
𝜎𝒙∗ 𝜎𝒙
∗
𝝈 = 𝜎𝒚 = 𝑯𝑇 𝜎𝒚
∗
(23)
∗
𝜎𝒙𝒚 𝜎𝒙𝒚
and the additional geometric stiffness matrix 𝑲∗∗ 𝜎 takes the following form:

𝑲∗∗
𝜎 = 𝑩𝑇 𝑷∗∗
𝟎𝒊 𝑩 𝑑𝑉0 (24)
𝑉0
∗∗
where 𝑷𝟎𝒊 is obtained from:
𝛿𝑯𝑇 𝝈 = 𝑷∗∗ ∗∗
𝟎𝒊 𝛿𝜺0 = 𝑷𝟎𝒊 𝑩𝛿𝒂0 (25)
Upon solving the incremental equilibrium equations for the displacement increments ∆ai0 and evaluating
the total displacements a incremental strains are obtained as:
1
∆𝜺𝑖0 = 𝑯 𝐁𝟎 + 𝐁𝑳 𝒂𝑖0 + 𝐁𝑳 ∆𝒂𝑖0 ∆𝒂𝑖0 (26)
2
The stress increments are then given by:
∆𝝈𝑖0 = 𝑫∆𝜺𝑖0 (27)
And the total stresses are:
𝝈𝑖+1
0 = 𝝈𝑖0 + ∆𝝈𝑖0 (28)
The residual forces, for a new displacement increment, are then equal to:

−𝝋𝒊+𝟏 = 𝐑 − 𝑩𝑇 𝑯𝑇 𝝈𝒊+𝟏
𝟎 𝑑𝑉0
𝑉0

=𝐑− 𝑩∗𝑇 𝝈𝒊+𝟏

𝟎 𝑑𝑉0 (29)
𝑉0

III. NUMERICAL RESULTS ANDDISCUSSION

The finite element TLG and TLE formulations described in the above section were implemented in the
NUSAP coded by FORTRAN. Three numerical examples of large deformation problems were examined to
demonstrate the degree of accuracy that can be obtained by using the geometrically non-linear formulations based
on 4-node isoparametric plane stress/strain elements. The examples were also solved by using Green strains and
the approximate Engineering strains [15]. The results of the stresses of the three TL formulations are compared
with the true Cauchy stresses based on Logarithmic strains obtained using the formulation presented in references
[13] and [17].

3.1 Cantilever under point load at free end:

The TLE (Eng Exact), approximate TLE (Eng) and TLG (Green) formulations were tested by
analyzing the cantilever plate with vertical load at the free end. The cantilever is of dimensions L=2.5 m, D = 1
m and t = 0.1 m as shown in Figure (1). The numerical values of material property parameters are; Young's
modulus, E = 2x108 kN/m2 and Poisson’s ratio υ = 0.3. The structure is modeled with 40 equal size
isoparametric elements. The cantilever was, also, analyzed using Logarithmic (Log) strains for comparison.
Graphical comparison of results of the stresses at the support and at mid- span are presented in Figures
(2) to (7). Tables (1), (2) and (3) show the stresses at mid-span. The results of the true Cauchy stresses are also
shown. At mid-span the results for the direct stresses are almost identical for the Log and Eng exact
formulations. The slight differences may be attributed to the large strain value at mid-span. The Green values
vary greatly from the correct values and are almost zero in the x-direction for maximum load. The differences in
the shear stress values are mainly due to the assumption that the shear strain is small in the formulations other
than Eng Exact. The results for the direct stresses at the support are in close agreement. Those for the shear
stresses at the support are in close agreement for Eng Exact and Log formulations. The shear values from the
Eng formulation are approximate. The shear values from the Green formulation differ largely from the correct
values as expected.

www.ijeijournal.com Page | 19
Evaluation of Engineering Stresses as the “Correct” Measure of “Physical” Stresses in Large Strain

y P/2

D=1m x

L = 2.5 m

P/2
Fig. (1) Cantilever plate with vertical load at free end

Table (1): Average Nodal Stress in x-direction at mid span

Stress (kN/m2) Stress (kN/m2)
LOAD LOAD
(kN) Eng (kN)
Eng
Green Eng Log Green Eng Log
Exact Exact
0 0 0 0 0 92000 -5.64E+06 -5.84E+06 -6.63E+06 -4.86E+06
4000 -2.99E+05 -2.99E+05 -2.99E+05 -2.99E+05 100000 -5.82E+06 -5.83E+06 -6.65E+06 -4.51E+06
20000 -1.48E+06 -1.50E+06 -1.51E+06 -1.49E+06 116000 -5.89E+06 -5.09E+06 -5.70E+06 -2.82E+06
36000 -2.62E+06 -2.71E+06 -2.80E+06 -2.66E+06 132000 -5.47E+06 -3.03E+06 -2.79E+06 6.57E+05
52000 -3.69E+06 -3.89E+06 -4.14E+06 -3.75E+06 148000 -4.42E+06 8.10E+05 2.94E+06 6.54E+06
68000 -4.63E+06 -4.94E+06 -5.41E+06 -4.59E+06 164000 -2.62E+06 6.94E+06 1.24E+07 1.55E+07
84000 -5.37E+06 -5.67E+06 -6.38E+06 -4.95E+06 180000 8.12E+04 1.58E+07 2.63E+07 2.81E+07

Average Nodal Stress in x-direction at mid span Average Nodal Stress in y-direction at mid span
30000000
50000000
Green Green
25000000 45000000
Eng Eng
40000000
20000000 Log Log
35000000
Stress (kN/m2)

Eng Exact
Stress (kN/m2)

Eng Exact
15000000 30000000
25000000
10000000
20000000
5000000 15000000
10000000
0
0 25000 50000 75000 100000 125000 150000 175000 200000 5000000
-5000000
0

-10000000 -5000000 0 25000 50000 75000 100000 125000 150000 175000 200000

Load (kN) Load (kN)

Fig. (2): Average Nodal Stress in x-direction at mid span Fig. (3): Average Nodal Stress in y-direction at mid span

Table (2): Average Nodal Stress in y-direction at mid span

LOA Stress (kN/m2) Stress (kN/m2)
LOAD
D Eng Eng
Green Eng Log (kN) Green Eng Log
(kN) Exact Exact
0 0 0 0 0 92000 3.74E+05 9.81E+05 1.01E+06 1.99E+06
4000 2.42E+04 2.44E+04 2.44E+04 2.43E+04 100000 6.92E+05 1.68E+06 1.90E+06 3.02E+06
20000 1.20E+05 1.28E+05 1.38E+05 1.20E+05 116000 1.63E+06 3.81E+06 4.82E+06 6.08E+06
36000 1.98E+05 2.27E+05 2.77E+05 1.77E+05 132000 3.08E+06 7.24E+06 9.81E+06 1.09E+07
52000 2.22E+05 2.53E+05 3.65E+05 9.64E+04 148000 5.18E+06 1.24E+07 1.76E+07 1.80E+07
68000 1.35E+05 7.97E+04 2.34E+05 2.89E+05 164000 8.06E+06 1.96E+07 2.88E+07 2.79E+07
84000 1.39E+05 4.81E+05 3.91E+05 1.22E+06 180000 1.19E+07 2.93E+07 4.42E+07 4.11E+07

www.ijeijournal.com Page | 20
Evaluation of Engineering Stresses as the “Correct” Measure of “Physical” Stresses in Large Strain

Table (3): Average Shear Stress at mid span

Stress (kN/m2) Stress (kN/m2)
LOAD LOAD
Eng Eng
(kN) Green Eng Log (kN) Green Eng Log
Exact Exact
0 0 0 0 0 92000 -5.71E+05 -8.71E+05 -9.87E+05 -9.32E+05
4000 2.17E+04 -2.17E+04 -2.17E+04 -2.17E+04 100000 -6.48E+05 -1.05E+06 -1.22E+06 -1.12E+06
20000 -1.08E+05 -1.10E+05 -1.10E+05 -1.11E+05 116000 -8.31E+05 -1.48E+06 -1.83E+06 -1.54E+06
36000 -1.94E+05 -2.05E+05 -2.06E+05 -2.12E+05 132000 -1.06E+06 -1.99E+06 -2.53E+06 -1.98E+06
52000 -2.84E+05 -3.24E+05 -3.29E+05 -3.40E+05 148000 -1.35E+06 -2.48E+06 -3.08E+06 -2.25E+06
68000 -3.84E+05 -4.86E+05 -5.10E+05 -5.18E+05 164000 -1.72E+06 -2.73E+06 -2.90E+06 -2.06E+06
84000 -5.03E+05 -7.20E+05 -7.92E+05 -7.71E+05 180000 -2.17E+06 -2.42E+06 -1.07E+06 -9.74E+05

Average Shear Stress at mid span Average Nodal Stress in x-direction at support
500000

0 70000000
Green
0 50000 100000 150000 200000
-500000
60000000 Eng
Log
Stress (kN/m2)

50000000

Stress (kN/m2)
-1000000 Eng Exact
40000000
-1500000

30000000
-2000000
20000000
Green
-2500000
Eng. 10000000
-3000000 Log
Eng Exact 0
-3500000 0 25000 50000 75000 100000 125000 150000 175000 200000

Load (kN) Load (kN)

Fig. (4) Average shear stress at mid span Vertex Fig. (5) Average Nodal Stress in x-direction at support

Average Nodal Stress in y- direction at support Average Shear Stress at support

25000000 45000000
Green Green
40000000
Eng Eng
20000000 35000000 Log
Log
Stress (kN/m2))

Eng Exact
Stress (kN/m2)

Eng Exact 30000000

15000000
25000000

20000000
10000000
15000000

5000000 10000000

5000000

0 0
0 25000 50000 75000 100000 125000 150000 175000 200000 0 25000 50000 75000 100000 125000 150000 175000 200000

Load (kN) Load (kN)

Fig. (6) Average Nodal Stress in y-direction at mid span Fig. (7) Average shear stress at Support

3.2 Cantilever under pure bending at free end:

A cantilever subjected to pure moment is considered. The cantilever is of dimensions L = 3000 mm,
D = 300 mm and thickness t = 60 mm as shown in Figure (8). The numerical values of material property
parameters are Young's modulus, E = 210 GPa, and Poisson’s ratio,  = 0.3. The structure is modeled with a
mesh of 40-isoparametric elements. The mesh is of equal elements size of 150x150 mm. The variations in the
stresses at the support and at mid-span with load increments as computed by Eng exact formulation are
compared with the Eng and Green formulations from Ref [15] and the Log formulation result presented in Ref
[17]. The results are presented in Figures (9) to (14) and tables (4) to (9).
As can be seen from the tables the values of the stresses are generally small. The stresses in the x-
direction are in close agreement for all formulations up to the 24000 N load. The Eng Exact and Log values
clearly agree for all loads. There is a large difference between these values and the Green value for maximum
load. The stresses in the y-direction and the shear stresses at mid-span show a similar trend with the Green
values almost zero for all loads. This shows that the Engineering stresses, in contrast to the stresses obtained
using Green strains, are correct measures of the physical stresses. The stresses at the support vary linearly and
are all in close agreement. This is mainly due to the small strain values at the support.

www.ijeijournal.com Page | 21
 Evaluation of Engineering Stresses as the “Correct” Measure of “Physical” Stresses in Large Strain

P
D = 0.3 m x
P
L=3m
Figure 8: Cantilever under pure bending

Table (4): Average Nodal Stress in x-direction at mid span

LOAD Stress (N/mm2)
(N) Log Green Eng (Geom) Eng Exact
0 0 0 0 0
6000 1.24E+00 1.22E+00 1.24E+00 1.23E+00
12000 2.50E+00 2.41E+00 2.47E+00 2.43E+00
18000 3.58E+00 3.49E+00 3.54E+00 3.37E+00
24000 3.94E+00 4.35E+00 4.09E+00 3.63E+00
30000 2.71E+00 4.84E+00 3.58E+00 2.61E+00

Average Nodal Stress in x-direction at mid span Average Nodal Stress in y-direction at mid span
6 6
Eng Excat Eng Exac
5 Geom 5 Geom
Green Green
Stress (N/mm2)

4
Stress (N/mm2 )

4 Log Log
3
3
2
2
1

1 0
0 5000 10000 15000 20000 25000 30000 35000
0 -1
0 5000 10000 15000 20000 25000 30000 35000
Load (N)
Load (N)

Fig. (9): Average Nodal Stress in x-direction at mid span Fig. (10): Average Nodal Stress in y-direction at mid span

Table (5): Average Nodal Stress in y-direction at mid span

LOAD Stress (N/mm2)
(N) Log Green Eng (Geom) Eng Exact
0 0 0 0 0
6000 9.62E-02 9.75E-02 9.59E-02 9.13E-02
12000 1.05E-01 1.66E-01 1.21E-01 7.57E-02
18000 2.86E-01 1.33E-01 1.50E-01 3.18E-01
24000 1.67E+00 1.15E-01 1.10E+00 1.54E+00
30000 4.92E+00 7.18E-01 3.26E+00 4.18E+00

Table (6): Average Shear Stress at mid span

LOAD Stress (N/mm2)
(N) Log Green Eng (Geom) Eng Exact
0 0 0 0 0
6000 4.47E-05 6.77E-04 1.26E-03 2.67E-03
12000 1.50E-02 5.56E-03 1.48E-02 2.23E-02
18000 8.82E-02 2.32E-02 6.87E-02 8.97E-02
24000 2.83E-01 6.56E-02 2.02E-01 2.48E-01
30000 6.39E-01 1.47E-01 4.43E-01 5.40E-01

www.ijeijournal.com Page | 22
Evaluation of Engineering Stresses as the “Correct” Measure of “Physical” Stresses in Large Strain

Average Shear Stress at mid span Average Nodal Stress in x-direction at support
0.7 10
Eng Exac Eng Excat
9
0.6 Geom
Geom
8
0.5 Green Green

Stress (N/mm2 )
7
Stress (N/mm2)

Log Log
0.4 6

0.3 5
4
0.2
3
0.1
2
0 1
0 5000 10000 15000 20000 25000 30000 35000
-0.1 0
0 5000 10000 15000 20000 25000 30000 35000
Load (N)
Load (N)

Fig. (11) Average shear stress at mid-span Fig. (12) Average Nodal Stress in x-direction at support

Table (7): Average Nodal Stress in x-direction at support

LOAD Stress (N/mm2)
(N) Log Green Eng (Geom) Eng Exact
0 0 0 0 0
6000 1.78E+00 1.81E+00 1.80E+00 1.82E+00
12000 3.50E+00 3.64E+00 3.58E+00 3.66E+00
18000 5.19E+00 5.47E+00 5.35E+00 5.54E+00
24000 6.86E+00 7.31E+00 7.12E+00 7.44E+00
30000 8.52E+00 9.15E+00 8.90E+00 9.36E+00

Table (8): Average Nodal Stress in y direction at support

LOAD Stress (N/mm2)
(N) Log Green Eng (Geom) Eng Exact
0 0 0 0 0
6000 9.62E-02 9.75E-02 9.59E-02 9.13E-02
12000 1.05E-01 1.66E-01 1.21E-01 7.57E-02
18000 2.86E-01 1.33E-01 1.50E-01 3.18E-01
24000 1.67E+00 1.15E-01 1.10E+00 1.54E+00
30000 4.92E+00 7.18E-01 3.26E+00 4.18E+00

Average Nodal Stress in y-direction at support Average Shear Stress at support

7
3.5 Eng Exac
Eng Exac 6 Geom
3 Geom
Green
Green 5
Stress (N/mm2)

2.5
Stress (N/mm2)

Log
Log
4
2

1.5
3

1 2

0.5 1

0 0
0 5000 10000 15000 20000 25000 30000 35000 0 5000 10000 15000 20000 25000 30000 35000
Load (N) Load (N)

Fig. (13): Average Nodal Stress in y-direction at support Fig. (14): Average Shear Stress at support

Table (9): Average Shear Stress at support

LOAD Stress (N/mm2)
(N) Log Green Eng (Geom) Eng Exact
0 0 0 0 0
6000 1.23E+00 1.23E+00 1.25E+00 1.24E+00
12000 2.44E+00 2.43E+00 2.52E+00 2.50E+00
18000 3.64E+00 3.61E+00 3.82E+00 3.77E+00
24000 4.81E+00 4.77E+00 5.14E+00 5.05E+00
30000 5.97E+00 5.90E+00 6.49E+00 6.36E+00

www.ijeijournal.com Page | 23
Evaluation of Engineering Stresses as the “Correct” Measure of “Physical” Stresses in Large Strain

3.3 Clamped beam under point force

A beam with two-fixed ends is considered. The beam is of length L = 200 mm, height D = 10 mm and
thickness 1 mm as shown in Figure (15). The numerical values for material property parameters are Young's
modulus, E = 210 GPa, Poisson's ratio, = 0.3. The beam is modeled with a mesh of 20-elementes.
The variation of the stresses at the support and at mid-span with the load increments as computed from
the Eng Exact formulation and for the approximate Eng and Green formulations (Ref. [15]), are compared with
the true stresses(Ref. [17]) in Figures (16) to (21).
y P

D = 10 mm x

L = 200 mm

Fig. (15): Clamped beam under point force

Tables (10), (11) and (12) show the values for average nodal stresses at mid-span. Very large loads
were applied in example resulting in large strains.
There is a marked difference between the Green and the other formulations’ values for the stresses at
the support for large load values. This is expected for cases of large strain. The Eng Exact values are the closest
to the Log (Cauchy) values for direct stresses. A similar trend is shown by the values of the stresses in the x-
direction at mid span with a maximum percentage difference between the Eng Exact and Log values of about
7% (around 69% for the Green and 37% for Eng). The stress at mid-span in the y-direction shows a similar
variation with the Eng Exact and Log values in close agreement and continuously increasing and the Green
values almost constant and close to zero. The maximum difference between the Eng Exact values and Eng
values is around 45%. This clearly shows the effect of assuming that the shear strain is small. The Eng Exact
shear stresses at mid-span values are large, compared to the other values, and are almost of a linear variation.
The Log and Eng shear values are in close agreement. This results from the assumption that the shear strain is
small in these formulations. The Green shear values at mid-span are very small compared to the values from the
other formulations and are of a non- uniform alternating nature. Thus, the Green strain formulation is not
suitable for evaluating the correct physical stresses. Also, the assumption that the shear strain is small limits the
use of the approximate Eng formulation for cases of small strain. Table (13) and Figure (22) show the maximum
principal stresses at mid-span for the Log and Eng Exact formulations. These are almost identical with a
maximum difference of about 6%. Hence, the stresses obtained using the Eng Exact formulation are considered
to be the correct measure of the physical stresses in large strain GNL.

Average Nodal Stress in x-direction at support Average Nodal Stress in y-direction at support

9000
Eng Exact 3000
8000 Eng Exact
Geom
2500 Geom
7000 Green
Green
Stress (N/mm2 )

Stress (N/mm2)

6000 Log
2000 Log
5000

4000 1500

3000
1000
2000
1000 500

0
0
0 5000 10000 15000 20000 25000 30000 35000 40000
0 5000 10000 15000 20000 25000 30000 35000 40000
Load (N)
Load (N)

Fig. (16): Average Nodal Stress in x-direction at support Fig. (17): Average Nodal Stress in y-direction at support

www.ijeijournal.com Page | 24
Evaluation of Engineering Stresses as the “Correct” Measure of “Physical” Stresses in Large Strain

Average Shear Stress at support Average Nodal Stress in x-direction at mid span
5000 25000
Eng Exact Eng Exact
4500
Geom Geom
4000 Green 20000 Green
3500 Log Log

Stress (N/mm2 )
Stress (N/mm2 )

3000 15000

2500

2000 10000

1500

1000 5000

500

0 0
0 5000 10000 15000 20000 25000 30000 35000 40000 0 5000 10000 15000 20000 25000 30000 35000 40000

Load (N) Load (N)

Fig. (18): Average Shear Stress at support Fig. (19): Average Nodal Stress in x-direction at mid span

Table (10): Average Nodal Stress in x-direction at mid span

Stress (N/mm2) Stress (N/mm2)

LOAD LOAD
(N) Eng (N) Eng Eng
Log Green Eng Exact Log Green
(Geom) (Geom) Exact
0 0 0 0 0 17800 7.47E+03 5.31E+03 6.64E+03 7.52E+03
1000 1.25E+03 1.23E+03 1.23E+03 1.23E+03 19400 8.40E+03 5.60E+03 7.30E+03 8.50E+03
2600 3.46E+03 3.28E+03 3.26E+03 3.31E+03 21000 9.44E+03 5.86E+03 7.98E+03 9.55E+03
4200 4.23E+03 4.00E+03 3.92E+03 4.10E+03 22600 1.06E+04 6.10E+03 8.69E+03 1.07E+04
5800 4.15E+03 3.98E+03 3.84E+03 4.13E+03 24200 1.18E+04 6.30E+03 9.41E+03 1.19E+04
7400 4.07E+03 3.95E+03 3.77E+03 4.14E+03 25800 1.31E+04 6.48E+03 1.01E+04 1.31E+04
9000 4.18E+03 4.07E+03 3.83E+03 4.29E+03 27400 1.45E+04 6.62E+03 1.08E+04 1.43E+04
10600 4.47E+03 4.34E+03 4.01E+03 4.59E+03 29000 1.60E+04 6.74E+03 1.15E+04 1.56E+04
12200 4.93E+03 4.71E+03 4.26E+03 5.03E+03 30600 1.75E+04 6.82E+03 1.22E+04 1.69E+04
13800 5.52E+03 5.18E+03 4.55E+03 5.59E+03 32200 1.91E+04 6.88E+03 1.29E+04 1.82E+04
15400 6.24E+03 5.72E+03 4.86E+03 6.26E+03 33800 2.07E+04 6.92E+03 1.35E+04 1.95E+04
17000 7.07E+03 6.32E+03 5.16E+03 7.04E+03 35400 2.23E+04 6.93E+03 1.41E+04 2.07E+04

Table (11): Average Nodal Stress in y-direction at mid span

Stress (N/mm2) Stress (N/mm2)
LOAD LOAD
(N) Eng (N) Eng Eng
Log Green Eng Exact Log Green
(Geom) (Geom) Exact
0 0 0 0 0 17800 1.76E+03 6.18E+02 1.57E+03 2.20E+03
1000 3.21E+02 3.17E+02 3.21E+02 3.39E+02 19400 2.35E+03 7.55E+02 2.01E+03 2.90E+03
2600 7.52E+02 7.14E+02 7.41E+02 8.76E+02 21000 3.03E+03 8.75E+02 2.48E+03 3.68E+03
4200 6.43E+02 5.45E+02 6.08E+02 7.71E+02 22600 3.80E+03 9.77E+02 2.97E+03 4.54E+03
5800 3.39E+02 2.13E+02 2.97E+02 3.94E+02 24200 4.65E+03 1.06E+03 3.48E+03 5.45E+03
7400 1.52E+02 1.93E+01 1.20E+02 1.58E+02 25800 5.58E+03 1.12E+03 4.00E+03 6.41E+03
9000 1.13E+02 2.03E+01 9.79E+01 1.10E+02 27400 6.58E+03 1.15E+03 4.51E+03 7.40E+03
10600 1.98E+02 1.07E+01 1.98E+02 2.23E+02 29000 7.65E+03 1.17E+03 5.03E+03 8.43E+03
12200 3.87E+02 1.13E+02 3.92E+02 4.69E+02 30600 8.77E+03 1.17E+03 5.54E+03 9.47E+03
13800 6.69E+02 2.47E+02 6.60E+02 8.33E+02 32200 9.93E+03 1.15E+03 6.03E+03 1.05E+04
15400 1.04E+03 3.95E+02 9.89E+02 1.30E+03 33800 1.11E+04 1.11E+03 6.50E+03 1.16E+04
17000 1.50E+03 5.45E+02 1.37E+03 1.88E+03 35400 1.23E+04 1.06E+03 6.96E+03 1.26E+04

Average Nodal Stress in y-direction at mid span Average Shear Stress at mid
14000 1400
Eng Exact Eng Exact
12000 Geom 1200 Geom
Green
10000 1000 Green
Stress (N/mm2)

Log
Stress(N/mm2 )

8000 800 Log

6000 600

4000 400

2000 200

0 0
0 5000 10000 15000 20000 25000 30000 35000 40000 0 5000 10000 15000 20000 25000 30000 35000 40000
-200
Load (N)
Load(N)

Fig. (20): Average Nodal Stress in y-direction at mid span Fig. (21): Average Shear Stress at mid span
www.ijeijournal.com Page | 25
Evaluation of Engineering Stresses as the “Correct” Measure of “Physical” Stresses in Large Strain

Table (12): Average Shear Stress at mid span

Stress (N/mm2) Stress (N/mm2)
LOAD LOAD
Eng Eng Eng
(N) Log Green Eng Exact (N) Log Green
(Geom) (Geom) Exact
0 0 0 0 0 17800 2.42E+02 1.26E+02 2.45E+02 5.14E+02
1000 3.73E-01 1.12E+00 1.43E+00 1.87E+00 19400 2.70E+02 1.18E+02 2.63E+02 5.82E+02
2600 2.11E+01 1.31E+01 1.07E+01 3.90E+01 21000 3.01E+02 1.08E+02 2.82E+02 6.52E+02
4200 1.12E+01 2.21E+00 8.19E+00 2.74E+01 22600 3.33E+02 9.55E+01 3.01E+02 7.23E+02
5800 2.74E+01 4.11E+01 5.19E+01 3.50E+01 24200 3.66E+02 8.01E+01 3.21E+02 7.94E+02
7400 6.55E+01 7.47E+01 9.21E+01 1.03E+02 25800 4.02E+02 6.24E+01 3.42E+02 8.66E+02
9000 9.75E+01 9.85E+01 1.25E+02 1.66E+02 27400 4.38E+02 4.24E+01 3.63E+02 9.38E+02
10600 1.25E+02 1.14E+02 1.52E+02 2.27E+02 29000 4.76E+02 2.04E+01 3.87E+02 1.01E+03
12200 1.51E+02 1.24E+02 1.75E+02 2.88E+02 30600 5.16E+02 3.49E+00 4.12E+02 1.08E+03
13800 1.76E+02 1.29E+02 1.96E+02 3.50E+02 32200 5.57E+02 2.92E+01 4.40E+02 1.16E+03
15400 2.01E+02 1.31E+02 2.16E+02 4.14E+02 33800 5.99E+02 5.66E+01 4.70E+02 1.23E+03
17000 2.28E+02 1.28E+02 2.35E+02 4.80E+02 35400 6.43E+02 8.55E+01 5.02E+02 1.31E+03

Table (13) Maximum Principal Stress at Mid-span

Stress (N/mm2) Stress (N/mm2) Stress (N/mm2)
LOAD LOAD
Eng LOAD Eng Eng
(N) Log Log (N) Log
Exact Exact Exact
0 0.0 0.0 12200 4935.0 5048.0 24200 11819.0 11996.0
1000 1250.0 1230.0 13800 5526.0 5616.0 25800 13121.0 13210.0
2600 3460.0 3311.0 15400 6248.0 6294.0 27400 14524.0 14425.0
4200 4230.0 4100.0 17000 7079.0 7084.0 29000 16027.0 15740.0
5800 4150.0 4130.0 18600 7931.0 8054.0 30600 17530.0 17054.0
7400 4071.0 4143.0 20200 8923.0 9076.0 32200 19134.0 18371.0
9000 4182.0 4297.0 21800 10015.0 10178.0 33800 20737.0 19687.0
10600 4474.0 4602.0 23400 11217.0 11390.0 35400 22341.0 20907.0

Max. Principal stress at mid-span

25000
Log
20000 Eng Exact
Stress (N/mm2)

15000

10000

5000

0
0 5000 10000 15000 20000 25000 30000 35000 40000

Load (N)

Fig. (22) Maximum Principal Stress at Mid-span

IV. CONCLUSIONS
Based on the results of the numerical examples, it can be concluded that:
1. The Total Lagrangian solutions based on the Green strains and 2 ndPiola-Kirchhoff stresses while being not
suitable for evaluating the correct physical stresses are necessary for use as a base for the solutions based on
Engineering stresses.
2. The exact Engineering strain based formulation results in correct physical stresses, which are very close to
the true Cauchy stresses especially for small and moderately large strains. Bearing in mind the fact that the
elastic constants are evaluated using these stresses, these formulations must be used when stresses are
required in a Total Lagrangian analysis.
3. The formulation based on approximate Engineering strains give excellent results in structures wherein the
shear strains are small.
www.ijeijournal.com Page | 26
Evaluation of Engineering Stresses as the “Correct” Measure of “Physical” Stresses in Large Strain

4. The use of Logarithmic strains is necessary when the exact true stresses are required. The results from the
Log formulation presented here can be further enhanced by removing the restriction of small shear strains.
5. The formulation based on the exact Engineering strains can be easily extended to three-dimensional
analysis.

REFRENCES
[1] Yang, Y. B. and Kuo, S. R., Theory and Analysis of Nonlinear Framed Structures, Prentice Hall, Simon & Schuster (Asia), 1998,
Singapore.
[2] Crisfield, M. A., Non-linear Finite Element Analysis of Solids and Structures, Volume 1, John Wiley & Sons Ltd., August 1997,
Chichester, England.
[3] Zienkiewicz O. C. and Taylor, R. L., The Finite Element Method for Solids and Structural Mechanics, 6th edition , Butterworth
Heinemann 2005, Elsevier.
[4] Mohamed, A. E., A Small Strain Large Rotation Theory and Finite Element Formulation of Thin Curved Beams, Ph.D. Thesis, 1983,
The City University, London.
[5] Belytschko, T., Finite Elements for Nonlinear Continua & Structures, North-Westren University, 1998, Evanston.
[6] Marinkovic`, D., Koppe, H. and Gabbert, U., Degenerated shell element for geometrically nonlinear analysis of thin-walled
piezoelectric active structures, Smart Mater. Struct. 17, 2008, 015030 (10pp), IOP Publications, G.B.
[7] Wood, R. D. and Zienkiewicz, O. C., Geometrically Nonlinear Finite Element Analysis of Beams, Frames, Arches and Axisymmetric
Shells, Computers & Structures, 1977, Vol.7, pp 725-735, PergamonPress, G.B.
[8] Surana, K. S. and Sorem, R. M., Geometrically Nonlinear Formulation for Three Dimensional Curved Beam Elements with Large
Rotations, Int. J. Num. Meth. Engng., 1989, Vol.28, 43-73.
[9] Djermane, M., Chelghoum, A. ,Amieur, B. and Labbaci, B., Linear and Nonlinear Thin Shell Analysis using a Mixed Finite with
Drilling Degrees of Freedom, Int. J. Applied Engng. Research, 2006, Vol. 1 No. (2) pp 217-236.
[10] Bonet, J. and Wood, R.D., Nonlinear Continuum Mechanics for Finite Element Analysis, CambridgeUniversity Press, 1997,
Cambridge, U.K.
[11] Mohamed, A. E. and Adam, F. M., Large Deformation Finite Element Analysis of Shell Structures, Journal of Science &
Technology, June 2003, Vol.4-No. 2, 47-58, SUST, Khartoum, Sudan.
[12] Adam, F. M. and Elzubair, A., Large Deformation Finite Element Analysis of Shells, Degenerated Eight Nodes Shell Element, LAP
LAMBERT Academic Publishing, 2012, Germany.
[13] Akasha Hilal, N. M., Development of Geometrically Nonlinear Finite Element Program using Plane Stress/Strain Elements based on
Engineering and True Stress Measures, Ph. D. Thesis, SUST, October 2009, Khartoum, Sudan.
[14] Akasha, N. M. and Mohamed, A. E., Geometrically Nonlinear Analysis using Plane Stress/Strain Elements based on Alternative
Strain Measures”, Journal of Science & Technology; Engng. Comp. Sciences, June 2012, Vol. 13- No. 1- 1-12, SUST, Khartoum,
Sudan.
[15] Akasha, N. M. and Mohamed, A. E., Evaluation of Engineering Stress for Geometrically Nonlinear Plane Stress/Strain Problems,
JASER, June 2012, Vol. 2, No. 2, , 115-125, Design for Scientific Renaissance.
[16] Greco, M. and Ferreira, I. P., Logarithmic strain measure applied to the nonlinear formulation for space truss analysis, Finite
Elements in Analysis and Design, 2009, 45, 632-639.
[17] Akasha, N. M. and Mohamed, A. E., Evaluation of True Stress for Geometrically Nonlinear Plane Stress/Strain Problems, JASER,
March 2012, Vol.2, No.1, 68-79, Design for Scientific Renaissance.

www.ijeijournal.com Page | 27